The use of kernel principal component analysis and discrete wavelet transform to determine the gas and oil interface

Abstract

The variations of pore fluid energy encoded within resistivity well logs can be considered as a significant attribute in the determination of reservoir fluid contacts. As a paramount technique in isolation and manipulation of certain patterns hidden in masses of data, wavelet analysis can effectively unveil this attribute. In this study, the discrete wavelet transform was employed on new well logs generated by kernel principal component analysis to monitor the pore fluid energy of pay zones at two previously appraised wells. An expert wavelet-based model was extracted by revealing the latent pattern of pore fluid energy variations. This model was then used to specify the gas and oil interface in a target well contiguous with the appraised wells. The gas and oil interface obtained from the expert wavelet-based model was confirmed by the drill stem test analysis. Results of this investigation suggest that monitoring pore fluid energy with such a method can be considered a highly functional attribute in determining the gas and oil interface.